Hallo from Germany

[Horsty]

Hallo there!

my name is Jan.
I'm scrolling through the threads since a few days now and I really like what you're doing so far.

I have a Audi Coupé Quattro 20V (7A engine / 5-cylinder) which comes standard with a Hitachi (2-connector) engine control unit and i already upgraded it to the later 4-connector Hitachi version.

that gave me a better idle control valve, a real tps sensor (not just idle and full throttle switch), and i was so able to connect the newer hitachi ecu to the vw dignostics tool. -> vcds or vagcom

setup right now:
- tps sensor
- 2x knock sensor (one for cylinder 1,2,3 and one for cylinder 4,5)
- Ignision Distibutor with hallsensor
- 2x vr sensor at the flywheel (1x VR crankhome 1 teeth and 1x VR 135 crankteeth)
- 3-pin pwm idle control valve
- MAF
- 5x Bosch 315cc injection valves
- narrow band o2-sensor

the main problem is, that the idle is pretty shitty and you have to tune the engines idle rpm and AF-ratio with the engine running in diagnostic mode and some sort of o2-test equipment. On top of this in many times the engine dies in diagnostic mode. that whole thing is quite a pain in the ass.

so i want to bring it to some sort of this in the future:
- rusefi ecu
- tps sensor
- 2x knock sensor
- map sensor
- intake air temperature sensor
- baro correction (i read its not yet supported atm?!)
- crank sensor (60/2 if required)
- camsensor
- electrical gaspedal (i have a audi a4 b5 gaspedal laying around)
- electrical throttle body (i don't know which one yet)
- wide band o2-sensor

and i think the injectors can stay. they should be good for now.


i really like that the stuff you guys are inventing there is open source!!!

[AndreyB]

Welcome and thank you for the warm words

[Horsty]

So i just ordered this thing and i hope it's the correct one:
STM32F407G-DISC1

[kb1gtt]

Welcome along, sounds like an interesting project.

Have you found the wiki yet? Just incase you have not here's a link.
http://rusefi.com/wiki/index.php?title=Manual:Hardware_Frankenso_board

Is that natural aspirated and manifold injected? I'm also a Quattro driver. However mine is that TFSI 2.0 4 cyl. It's my daily and runs well, so I haven't touched it yet.

[Horsty]

thank you.

yes, I have seen the wiki already. I just wanted that someone confirms my order.

the 7A engine (thats the engine code of the 5-cylinder 20V non turbo) in my coup is naturaly aspirated and manifold injected.
i have some sort of plan to supercharge it, but that could take a while.

tfsi engines btw. are slightly crap.

[kb1gtt]

Sounds like rusEFI could work well for your engine. Do you know is if VR or HALL crank signals?

I agree there are short comings on the TFSI this entire gas direct injection is a maintenance problem. Diesel is a lubricant gas it a detergent, so hard to keep lubricated well. As well they are impossible to reach things when you need to make a repair. The trade off is more HP for less weight. I could get many more HP's if I simply change a bit or two in the ECU's brain.

[stefanst]

More Germans on here - Willkommen
Nice choice of car. I'd love to see some pictures.

Schoene Gruesse aus Amerika!

[Horsty]

@kb1gtt: the crank signals are both VR. the camsensor on the distributor is a hall sensor.

@stefanst: moin! sorry, but I'm not going to post any pictures of the car.

[kb1gtt]

Did you say it has 2 crank shaft sensors?

Do I recall that the cam shaft sensor in the distributor is just to verify cyl 1 has not wondered. Basically I think this pulse is to verify that someone did not try to adjust the timing by spinning the distributor like you would do on an older system.

Do you know your crank pulse pattern? The crank angle is decoded by a bit of code that can be changed via configuration parameter. There are many existing crank decoders to choose from, and if your wheel pattern does not work with the existing decoders, then a new decode will have to be made. Once a crank wheel decoder has been made for your particular crank pattern, you can select which crank pattern is used by rusEFI using software configuration. I believe the existing crank decoders have all used either 1 or 2 inputs signals. If yours has 3 input signals, that may cause some software problems. So lets learn what your signals are, such that we can figure out if yours can be done with 2 signals, or if it would require 3 signals.

[Horsty]

hi there,

your thoughts on that are totally right i guess.
and after having a look at the diagram again i would suggest that the second crank sensor is not really needed when the cam sensor is in the correct position.


these are the diagrams i found in the audi service manual.

1. cam hall-sensor (distributor)
2. crank 1
3. crank 2

sensor.JPG

sensor2.JPG

Some people remove both crank sensors and the rear crank seal (behind the flywheel) and replace the crankseal with one from the vw crafter or so.
That one has an inbuilt 60/2 sensor.

clk0e9fff0v8rxvp3.jpg

And i even found the ignition map for 95 octane fuel.

zündkennfeld 95 oktan.JPG

[kb1gtt]

I believe there is a 60-2 crank decoder already. So that and the single pulse from the CAM to tell if it is above 360 deg and 720deg should work reasonably well. Is the 60-2 wheel an option? That would likely make it easier for @russian, as it would likely require less software resources.

Cool find about the ignition advance.

[Horsty]

thank you for your help so far.

the 60/2 is an option the next time I tear that engine out of the bay again.
can i stick to the tripple sensor setup so far?


Btw. I got this like an hour ago.

IMG_9552.JPG

[AndreyB]

what is your plan for the mainboard? default frankenso has two VR channels, one can add two more with an expansion board.

[Horsty]

i seriously really don't know yet.

what i do know is that I'm going to build a completly new ecu housing with new mil-spec connectors like amphenol pt-series or so and a pretty much new wiring harness.
rayfast 55m wires would be nice i guess. so space should not be that big of deal.

apart from that, I'll probably gonna relocate the ecu from between the firewall and the interior ventilation housing to the passenger sides foot well or even behind the passenger seat.
does someone see any issues with that?

[kb1gtt]

I believe he has enough signals with the two below. I believe one is VR and pulses once every 120 degrees. Then the other one I believe is the Hall in the distributor, which provides a pulse once every 720deg, which allows you to know if you are + or - 360 deg. What I do not know is if the 120 degree pulse is 120 degrees. It appears it indicates TDC for each cyl, and that is commonly not evenly spaced.

[Horsty]

hallgebersignal = signal from the hall-sensor at the distributor
intern errechnetes Zylindersignal = internally calculated cylinder signal (calculated in the hitachi mpi, so I can't work with that)

[stefanst]

So the only signal we have to work with is every two revolutions cylinder 1 in TDC - correct?
That should be doable, but will lack some accuracy under highly dynamic conditions (accelerating hard in first gear).
The 135 pulse/ rotation signal is presumably an output from the ignition system to the tachometer (rpm-meter)?

Edit: Never mind- I missed the part with the 2 crank sensors.

[Horsty]

no problem.

stm board is up and running, connected to tuner studio and so on.

now i should make a serious decision about the further electonics...

IMG_9565.JPG

IMG_9563.JPG

[Horsty]

Look what i've found:

IMG_9567.JPG

IMG_9568.JPG

Audi 80 automatic ac connector which fits perfectly to the second gen. 7A stock engine wiring loom.

IMG_9575.JPG

Was quite hard to get that connector of that old bord, but the heatgun did the trick.

ordered parts:
- LT1084-CP5 5V/5A low dropout fixed voltage regulator
- some universal pcb's 150x200mm

[Horsty]

would this work?

Audi Trigger.JPG

[AndreyB]

- 2x vr sensor at the flywheel (1x VR crankhome 1 teeth and 1x VR 135 crankteeth)

does not the picture above show 1 tooth on CAM and 18 teeth on CRANK?

This would work
1) you would need 'custom skipped wheel, 18 total 0 skipped'
2) 4 stroke crank driven
3) VVT/sync on CAM wire, see http://rusefi.com/wiki/index.php?title=Manual:Software:VVT

[Horsty]

I'm sorry!!! the drawing is not accurate! it's 135 teeth on the crank.

just didn't want to draw that much.

[AndreyB]

I am suspecting same would work with total=1 and skipped=0 but hard to say without a cleaner picture.

Also while I am not aware of specific issues with so few teeth I wonder why so few, just two seems like a really, really low number which would give low trigger precision?

[Horsty]

got the red and orange led steady on the third time now after burning something into the stm via tunerstudio with the todays .ini and firmware loaded.
the only way back to the blinking blue is a full chip erase.

any ideas?

[AndreyB]

It does not like the configuration. Tuner Studio has a gauge for lastErrorCode and rusEfi console has a more detailed readable message.

[Horsty]

My dad milled some pieces out of the crank dampener pulley so I now have 60/2 trigger wheel.

IMG_9687.JPG

[kb1gtt]

That's great. 60-2 is a good wheel. I was a bit concerned with the 135 tooth wheel that you would have problems capturing the signal electrically. A 60-2 wheel should have minimal issues.

[Horsty]

i would rather say there are other issues now.

need to design a new vr-sensor bracket...
i'll probably integrate it in the torque support or something.

shot this thing on ebay by the way. ten bucks... freaking bargain if it works.

s-l1600.jpg

Datasheet

[puff]

Won't this cause sort of a disbalance because of these missed teeth? (at first I wanted to compensate it by drilling extra holes on the opposite side of the disk, but then I realised that I was removing some extra spacer which initially carried on balancing weights, and so I gave up that idea.

10$ including shipping?

Speaking of a mount, I used a jigsaw to cut two pieces our of aluminum electric busbar, and then went to a local tire/wheel rapair, they welded these pieces together for a buck.

[kb1gtt]

I've heard mixed reviews of the importance of balancing. Generally high rev and high performance does the balancing, and generic OEM does not. I understand that the extra balance will allow higher revs under higher load with out pushing through the oil and damaging the journal bearing, as well it makes for a smoother ride. However if you are not pushing the limits of the oil, and you don't mind some vibration from the engine, it's less important.

How does one get a wheel like this balanced? I see the holes that were drilled into it originally, probably from the MFG. Are there shops that can re-balance you balancer? Is there a DIY way to balance it?

BOSCH MAP looks fairly straight forward to used with Franekenso board. You might need to change a couple resistors on the Frankenso to get the MFG suggested bias resistance, but it looked like it shouldn't be a problem as we have locations for those resistors to be installed.